Scope

Using S&P500 data that I've pulled in a different python script, evaluate different features for correlation to stock price change, and use this for my equity trend dash. There is still more work to do here - additional features to look at, and something more elegant that the gradient boosted tree algo from SKLearn.

Import relevant data

Calculate change

This needs is over the total period available in the data, should look at period to period changes in the future.

Set categorical class

Look at correlations to price changes

Plot different variables with high correlation (positive or negative) against price change

Remove outliers and run correlations again

Use feature importance and compare to correlations

Try a few decision tree models for classification

Summary May 2022

Features to monitor for future stock selections based on above analysis include: